Skip to main content

numpy array with labeled axes

Project description

Title : axarray Author : sylvain.guieu@gmail.com

# Introduction axarray is a numpy array where axes can be labeled.

The idea is to be abble to manipulate array, and do operation on axis without knowing the array shape order but on knowing labels related to the ‘phisical’ meaning of the axes.

Often in science, it is usefull to name the array axes by an inteligible label. For instance, for 2d images taken at different time, axes name of the obtain cube could be [“time”, “y”, “x”]

axarray object aims to do that. For instance a.mean(axis=”time”) will execute the mean on the axis labeled “time” where ever it is.

Given a1 and a2, two axarray, binarry operation like a1+a2 can be performed even if the two axarray has different axes order as long as they have matching axis labels.

# Examples `python >>> a = axarray( np.random.random((10,4,5)), ["time", "y", "x"]) >>> b = a.transpose( ["x","time", "y"]) >>> b.axes ["x","time", "y"] `

can operate 2 transposed axarray as long as they match axis names `python >>> (a+b).axes ["time", "y", "x"] ` use the numpy frunction with axis labels `python >>> a.min(axis="time").shape (4,5) # similar to: >>> np.min(a , axis="time") `

axis can be alist of axis label `python >>> a.mean(axis=["x","y"]).shape (10,) ` one can use the conveniant apply method. Usefull in non-direct call as in a plot func for instance `python >>> a.apply(time_reduce=np.mean, y_idx=slice(0,2)).shape (2,5) `

transpose, reshape rename axes in one call `python >>> at = a.transform( [("pixel", "y","x"), "time"]) >>> at.shape (20, 10) # (4*5, 10) >>> at.axes ['pixel', 'time'] ` Extract a spectrum from image from named indices `python ### make some indices >>> iy, ix = axarray( np.indices( (3,4)), [0 ,"spatial", "freq"]) >>> ax[:,iy,ix].axes ['time', 'spatial', 'freq'] `

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

axarray-0.1.0.tar.gz (2.0 kB view hashes)

Uploaded Source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page